CN104461849A - Method for measuring power consumption of CPU (Central Processing Unit) and GPU (Graphics Processing Unit) software on mobile processor - Google Patents
Method for measuring power consumption of CPU (Central Processing Unit) and GPU (Graphics Processing Unit) software on mobile processor Download PDFInfo
- Publication number
- CN104461849A CN104461849A CN201410741891.6A CN201410741891A CN104461849A CN 104461849 A CN104461849 A CN 104461849A CN 201410741891 A CN201410741891 A CN 201410741891A CN 104461849 A CN104461849 A CN 104461849A
- Authority
- CN
- China
- Prior art keywords
- power consumption
- cpu
- measured
- program
- gpu
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The invention discloses a method for measuring power consumption of CPU (Central Processing Unit) and GPU (Graphics Processing Unit) software on a mobile processor. The method comprises the following steps: establishing a CPU power consumption model; modifying a program to be measured and recompiling; setting a platform to be measured; and operating the program to be measured and carrying out data processing. The method for measuring power consumption of a CPU and a GPU on the mobile processor, which is disclosed by the invention, effectively solves the problem that currently, a software developer difficultly and simultaneously acquires power consumption of the CPU and the GPU in the program executing process on a mobile processor of a mobile intelligent terminal. Any additional measurement tool is not required, the platform to be measured also does not need to be disassembled, power consumption of the CPU and the GPU on the mobile processor in the program executing process can be directly and accurately acquired on the mobile intelligent terminal and the developer of applications can be helped to design applications and games, which have low power consumption, for the intelligent terminal.
Description
Technical field
The present invention relates to merit measurement of power loss field, be specifically related to CPU and GPU software power consumption measuring method on a kind of mobile processor.
Background technology
Mobile calculation technique makes rapid progress, and the handheld terminal being representative with smart mobile phone, panel computer have also been obtained extensively universal and application.Along with the development of mobile calculation technique, particularly move the huge advance of processor technology, the range of application of these handheld terminals develops into from traditional communication, camera function the function such as high speed Internet access, 3D game, HD video that the smart mobile phone of today and panel computer generally have.
The mobile processor that hand-held intelligent terminal carries comprises CPU and GPU two parts usually, and CPU bears general computing power, and GPU bears the functions such as 2D, 3D graphic plotting, computing, display.The application program run on the mobile terminal such as smart mobile phone, panel computer at present, game, play the performance requirement of HD video to mobile processor more and more higher, bring larger power dissipation overhead simultaneously.Because the hand-held intelligent terminal such as smart mobile phone, panel computer is driven by lithium battery usually, the power dissipation overhead that therefore application program, game and video cause plays a part very crucial for Consumer's Experience.In order to develop the intelligent terminal application program of low-power consumption better, power consumption information when developer needs acquisition program to perform.But current existing Properties Analysis software only comparatively can obtain the execution time of application program, the power dissipation overhead moved when performing for program on CPU and GPU lacks effectively, accurate assay method.People expect to utilize power consumption measurement method, the execution time of flexible acquisition application program, power consumption, and then obtain consumption information, thus help developer to reduce the execution power consumption of program, promote user's experience.
In addition, because the component integrations such as CPU, GPU in most of mobile intelligent terminal are in same SoC chip, so cannot directly from the enterprising accommodating power of hardware and energy consumption measurement.
Summary of the invention
Goal of the invention: in order to solve the deficiencies in the prior art, directly carries out measurement of power loss by external program to CPU and GPU of mobile device, overcomes the problem that prior art lacks effectively, accurately measures method.
Technical scheme: CPU and GPU software power consumption measuring method on a kind of mobile processor, is characterized in that, the method comprise set up CPU power consumption model, revise program to be measured and recompilate, platform to be measured arrange, run program to be measured and data processing; The method comprises the following steps:
1) the cpu power model based on utilization rate is set up:
P＝ΣU
i×β
i
In formula:
P is the dynamic power of CPU;
U
iit is the utilization rate of i-th core cpu;
β
iit is the power consumption factor of i-th core cpu;
The intensive test set of moving calculation, the frequency of utilization of sample in operational process the battery momentary current magnitude of voltage that obtains from the operating system of platform to be measured and CPU record, by linear stipulations mode process the data obtained, obtain the β in cpu power model
i;
2) according to function and the power consumption test demand of each program module of platform to be measured, more than one code segment to be measured is divided; Add correlated performance at the head and the tail of each code segment to be measured and the head and the tail of whole program code and dissect code gtick (), and recompilate program to be measured;
3) operating procedure 2) the middle program to be measured recompilated: the program to be measured of this recompility records the moment performing each code segment and the moment executing each code segment in operational process; Meanwhile, the utilization rate of battery momentary current magnitude of voltage that sampling obtains from the operating system of platform to be measured and CPU is continued and record; The program to be measured of this recompility is dormancy a period of time before starting to perform and before terminating to perform, still lasting sampling battery momentary current magnitude of voltage and CPU usage between rest period; Dormancy is to make CPU/GPU be in Idle state a period of time, to measure background power consumption.
4) treatment step 3) the middle data recorded, specifically comprise the following steps:
4.1) instantaneous current value of each sampled point is multiplied with instantaneous voltage value, obtains the instantaneous power consumption values of each sampling instant; Getting instantaneous power consumption minimum value is background power consumption, and described background power consumption comprises the power consumption that each sampling instant CPU and GPU is in Idle state;
4.2) by the instantaneous power consumption values subtracting background power consumption of each sampled point, the actual total power consumption of each sampling instant CPU and GPU is obtained;
4.3) use step 1) in CPU power consumption model and step 3) in the CPU usage information of each sampling instant, by step 4.2) the actual total power consumption of CPU and GPU that obtains deducts step 1) in the power consumption that calculates of cpu power model formation, obtain the CPU power consumption of each sampling instant, and then the GPU power consumption of each sampling instant can be calculated;
4.4) by step 3) program to be measured that recompilates records the moment performing each code segment and the moment executing each code segment in operational process, calculates the execution time that each code segment expends; Utilize step 4.1) instantaneous power consumption values of each sampling instant that calculates, calculate the average power and total energy consumption that expend on each code segment.
Further, step 4) in, treat ranging sequence and take multiple measurements and average.
Beneficial effect:
On the mobile processor that the present invention proposes, the power consumption measurement method of CPU and GPU efficiently solves the power problems that software developer on current mobile intelligent terminal is difficult to move when obtaining program execution on CPU, GPU simultaneously.
The present invention proposes a kind of on mobile intelligent terminal, as smart mobile phone, panel computer, carry out the program execution time of mobile CPU and GPU, the measuring method of power consumption.
Utilize technical scheme of the present invention, without the need to any extra survey instrument, and without the need to disassembling platform to be measured, can power consumption that directly accurate acquisition program performs on mobile CPU and GPU on mobile intelligent terminal, Develop Application System personnel can be helped to design intelligent terminal application program and the game of low-power consumption.
The present invention can bring obvious benefit to needing the design application program of low-power consumption and the developer of hardware.
Accompanying drawing explanation
Fig. 1 is code flow schematic diagram
Fig. 2 is process flow diagram of the present invention
Fig. 3 is the output schematic diagram of execution time after 10 test data process, power consumption and energy consumption
Fig. 4 is the output schematic diagram of execution time after 20 test data process, power consumption and energy consumption
Embodiment
Below in conjunction with accompanying drawing the present invention done and further explain.
CPU and GPU software power consumption measuring method on a kind of mobile processor, is characterized in that, the method comprises to be set up CPU power consumption model, revise program to be measured and recompilate, platform to be measured setting, run program to be measured and data processing; The method comprises the following steps:
1) the cpu power model based on utilization rate is set up:
P＝ΣU
i×β
i
In formula:
P is the dynamic power of CPU;
U
iit is the utilization rate of i-th core cpu;
β
iit is the power consumption factor of i-th core cpu;
The intensive test set of moving calculation, the frequency of utilization of sample in operational process the battery momentary current magnitude of voltage that obtains from the operating system of platform to be measured and CPU record, by linear stipulations mode process the data obtained, obtain the β in cpu power model
i;
2) according to function and the power consumption test demand of each program module of platform to be measured, more than one code segment to be measured is divided; Add correlated performance at the head and the tail of each code segment to be measured and the head and the tail of whole program code and dissect code gtick (), and recompilate program to be measured;
3) operating procedure 2) the middle program to be measured recompilated: the program to be measured of this recompility records the moment performing each code segment and the moment executing each code segment in operational process; Meanwhile, the utilization rate of battery momentary current magnitude of voltage that sampling obtains from the operating system of platform to be measured and CPU is continued and record; The program to be measured of this recompility is before starting execution and terminate to perform front meeting dormancy a period of time, between rest period, still continue sampling battery momentary current magnitude of voltage and CPU usage;
4) treatment step 3) the middle data recorded, specifically comprise the following steps:
4.1) instantaneous current value of each sampled point is multiplied with instantaneous voltage value, obtains the instantaneous power consumption values of each sampling instant; Getting instantaneous power consumption minimum value is background power consumption, obtains the power consumption that each sampling instant CPU and GPU is in Idle state;
4.2) by the instantaneous power consumption values subtracting background power consumption of each sampled point, the actual total power consumption of each sampling instant CPU and GPU is obtained;
4.3) use step 1) in CPU power consumption model and step 3) in the CPU usage information of each sampling instant, extrapolate the CPU power consumption of each sampling instant, and then the GPU power consumption of each sampling instant can be calculated;
4.4) by step 3) program to be measured that recompilates records the moment performing each code segment and the moment executing each code segment in operational process, calculates the execution time that each code segment expends; Utilize step 4.1) instantaneous power consumption values of each sampling instant that calculates, calculate the average power and total energy consumption that expend on each code segment, treat ranging sequence and take multiple measurements and average.
As shown in Figure 1, carry out code segment division for treating ranging sequence and add at the head and the tail of each code segment and the head and the tail of whole program the schematic diagram that correlated performance dissects code.This sentences a kind of Heterogeneous Computing test procedure based on OpenCL is that example illustrates the present invention program, but this programme is not limited to OpenCL program.This test procedure contains the calculation task performed on CPU and GPU, and the function of foundation program and Properties Analysis demand have carried out code segment division.Test procedure can be divided into 4 code segments, and wherein A, D perform on CPU, and B, C perform on GPU.The program execution time that the present invention proposes, measurement of power loss are all divide based on code segment to carry out, thus final execution time and the power consumption results of property obtaining each code segment.
As shown in Figure 2, for measuring the process flow diagram operating in execution time of program on mobile processor CPU and GPU, power consumption and energy consuming process.This process flow diagram omits the process of CPU power consumption modeling.As we know from the figure, described measuring method by revise program to be measured and recompilate, platform to be measured arranges, runs program to be measured and the several part of data processing forms.First, need to treat ranging sequence modify and recompilate.Specifically, the beginning, the end that divide each functional module code segment obtained at code segment add gtick function; Gticki and gticke function is added respectively at the main function head and the tail place of program to be measured; If it is chronic that program is run, need to add gtickr and gtickp function in suitable place, often run a period of time just renewal measurement data daily record to make calling program.Gtick function is for recording the current moment, and the link is below for recording the execution time of each functional module.Gticki is initialization function, and be responsible for electric current, voltage sample sampler thread that startup linux kernel DLL (dynamic link library) provides, gticke stops function simultaneously.In addition, in order to measurement of power loss itself brings too much extra execution time and power dissipation overhead, the data that the present invention measures all are recorded in internal memory, and write external memory at the end of measuring or when running gtickp function.Meanwhile, program can complete Memory Allocation when initialization, can not occur realloc in measuring process, and the Memory Allocation power consumption brought to avoid it is to measuring the interference caused.During in order to measure, above-mentioned Insufficient memory deposits the situation of intermediate data, design introduces gtickr and gtickp two functions for realizing the time-out of electric current, voltage sample sampler thread and program to be measured and restarting, and in the gap suspended, data logging is being write external memory.Secondly, what need to treat lining platform carries out suitable setting, as fixing mobile processor CPU frequency, core number and display screen brightness, thus the uncertain factor of measurement of power loss of eliminating the effects of the act.Above-mentioned purpose can be realized by the initializtion script revising Android operation system.Next, testing results program on platform to be measured, obtains execution time and the power consumption results of property of mobile processor CPU and GPU simultaneously.As previously mentioned, gtick function is used for the current time that logging program performs, and gticki function is used for starting current, voltage sample sampler thread.In addition, in order to measure the power consumption of mobile processor accurately, for the program to be measured not using display screen, will close the display screen of platform to be measured, then this likely causes platform to be measured to enter sleep mode.For this problem, run Launcher program and can ensure test procedure always not by operating system kill, when ensureing that platform display screen to be measured is closed simultaneously, if program to be measured also off-duty terminate, then stop platform to be measured to enter sleep mode.In order to obtain more accurate execution time and measurement of power loss data, bench function is used for repeating to call program to be measured, and final data can be averaged as net result.Finally, the sampled data of execution moment of gtick function record and electric current, voltage is processed, draws out the performance table of execution time and power consumption, and calculate energy consumption result.After program end of run to be measured, need that the data logging recorded is sent back PC and add up, fig function is used for execution time, the calculating of power consumption and energy consumption, statistics and form and draws.
Concrete data processing method is as follows:
1) instantaneous current value of each sampled point is multiplied with instantaneous voltage value can obtains the instantaneous power consumption values of each sampling instant.
2) minimum value of getting the instantaneous power consumption of each sampling instant is background power consumption, and namely mobile platform CPU, GPU is in the power consumption of Idle state.By the instantaneous power consumption values subtracting background power consumption of each sampling instant, the total power consumption of each sampling instant CPU and GPU can be obtained.
3) by the CPU usage information of CPU power consumption model and each sampling instant, the CPU power consumption of each sampling instant can be extrapolated, utilize permit notification to deduct CPU power consumption number, and then calculate the GPU power consumption of each sampling instant.
4) execution of being recorded by step 3 to and execute moment of each code segment, the execution time that each code segment expends can be calculated.The instantaneous power consumption values of each sampling instant calculated on utilize, can calculate the average power and total energy consumption that expend on each code segment.
As shown in Figure 3, Figure 4, be respectively data processing after test procedure execution time on mobile processor CPU, GPU of obtaining, power consumption and energy consumption output schematic diagram.Data above dotted line are the overall performance of test procedure, and below is the Properties Analysis of selected parts some importance functional module (code segment).The overall performance of test procedure includes the execution time, measures power consumption and quiescent dissipation, consumption information, and the Properties Analysis of part of module includes the execution time of each several part, power consumption, energy consumption and proportion value.Mobile processor GPU measurement result shown in Fig. 4 further comprises GPU initialization (init), the data of CPU and GPU transmit operating performance results such as (mems).
The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the premise without departing from the principles of the invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.
Claims (2)
1. a CPU and GPU software power consumption measuring method on mobile processor, is characterized in that, the method comprise set up CPU power consumption model, revise program to be measured and recompilate, platform to be measured arrange, run program to be measured and data processing; The method comprises the following steps:
1) the cpu power model based on utilization rate is set up:
P＝ΣU
i×β
i
In formula:
P is the dynamic power of CPU;
U
iit is the utilization rate of i-th core cpu;
β
iit is the power consumption factor of i-th core cpu;
The intensive test set of moving calculation, the frequency of utilization of sample in operational process the battery momentary current magnitude of voltage that obtains from the operating system of platform to be measured and CPU record, by linear stipulations mode process the data obtained, obtain the β in cpu power model
i;
2) according to function and the power consumption test demand of each program module of platform to be measured, more than one code segment to be measured is divided; Add correlated performance at the head and the tail of each code segment to be measured and the head and the tail of whole program code and dissect code gtick (), and recompilate program to be measured;
3) operating procedure 2) the middle program to be measured recompilated: the program to be measured of this recompility records the moment performing each code segment and the moment executing each code segment in operational process; Meanwhile, the utilization rate of battery momentary current magnitude of voltage that sampling obtains from the operating system of platform to be measured and CPU is continued and record; The program to be measured of this recompility is dormancy a period of time before starting to perform and before terminating to perform, still lasting sampling battery momentary current magnitude of voltage and CPU usage between rest period;
4) treatment step 3) the middle data recorded, specifically comprise the following steps:
4.1) instantaneous current value of each sampled point is multiplied with instantaneous voltage value, obtains the instantaneous power consumption values of each sampling instant; Getting instantaneous power consumption minimum value is background power consumption, and described background power consumption comprises the power consumption that each sampling instant CPU and GPU is in Idle state;
4.2) by the instantaneous power consumption values subtracting background power consumption of each sampled point, the actual total power consumption of each sampling instant CPU and GPU is obtained;
4.3) use step 1) in CPU power consumption model and step 3) in the CPU usage information of each sampling instant, by step 4.2) the actual total power consumption of CPU and GPU that obtains deducts step 1) in the power consumption that calculates of cpu power model formation, obtain the CPU power consumption of each sampling instant, and then the GPU power consumption of each sampling instant can be calculated;
4.4) by step 3) program to be measured that recompilates records the moment performing each code segment and the moment executing each code segment in operational process, calculates the execution time that each code segment expends; Utilize step 4.1) instantaneous power consumption values of each sampling instant that calculates, calculate the average power and total energy consumption that expend on each code segment.
2. CPU and GPU software power consumption measuring method on a kind of mobile processor as claimed in claim 1, is characterized in that, step 4) in, treat ranging sequence and take multiple measurements and average.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410741891.6A CN104461849B (en) | 2014-12-08 | 2014-12-08 | CPU and GPU software power consumption measuring methods in a kind of mobile processor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410741891.6A CN104461849B (en) | 2014-12-08 | 2014-12-08 | CPU and GPU software power consumption measuring methods in a kind of mobile processor |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104461849A true CN104461849A (en) | 2015-03-25 |
| CN104461849B CN104461849B (en) | 2017-06-06 |
Family
ID=52907942
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410741891.6A Active CN104461849B (en) | 2014-12-08 | 2014-12-08 | CPU and GPU software power consumption measuring methods in a kind of mobile processor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104461849B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104778113A (en) * | 2015-04-10 | 2015-07-15 | 四川大学 | Method for correcting power sensor data |
| CN105446877A (en) * | 2015-11-04 | 2016-03-30 | 上海聚力传媒技术有限公司 | Method and apparatus for testing power consumption of mobile application |
| US20160274636A1 (en) * | 2015-03-16 | 2016-09-22 | Electronics And Telecommunications Research Institute | Gpu power measuring method of heterogeneous multi-core system |
| CN106686179A (en) * | 2016-12-27 | 2017-05-17 | 东南大学 | Mobile phone power consumption testing system and method based on test automation |
| CN109684144A (en) * | 2018-12-26 | 2019-04-26 | 郑州云海信息技术有限公司 | A kind of method and device of GPU-BOX system testing |
| CN111538636A (en) * | 2020-04-24 | 2020-08-14 | 深圳华锐金融技术股份有限公司 | Computer equipment determination method and device and storage medium |
| CN111782454A (en) * | 2020-08-05 | 2020-10-16 | 中国人民解放军国防科技大学 | Instruction EPI-based fine-grained GPDSP power consumption testing method, system and medium |
| CN112150631A (en) * | 2020-09-23 | 2020-12-29 | 浙江大学 | Real-time energy consumption optimization drawing method and device based on neural network |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130159741A1 (en) * | 2011-12-15 | 2013-06-20 | Travis T. Schluessler | Method, Apparatus, and System for Energy Efficiency and Energy Conservation Including Power and Performance Balancing Between Multiple Processing Elements and/or a Communication Bus |
| CN103279446A (en) * | 2013-06-09 | 2013-09-04 | 浪潮电子信息产业股份有限公司 | Isomerism mixed calculation multi-platform system using central processing unit (CPU)+graphic processing unit (GPU)+many integrated core (MIC) |
| CN104106053A (en) * | 2012-02-08 | 2014-10-15 | 英特尔公司 | Dynamic CPU GPU load balancing using power |
-
2014
- 2014-12-08 CN CN201410741891.6A patent/CN104461849B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130159741A1 (en) * | 2011-12-15 | 2013-06-20 | Travis T. Schluessler | Method, Apparatus, and System for Energy Efficiency and Energy Conservation Including Power and Performance Balancing Between Multiple Processing Elements and/or a Communication Bus |
| CN104106053A (en) * | 2012-02-08 | 2014-10-15 | 英特尔公司 | Dynamic CPU GPU load balancing using power |
| CN103279446A (en) * | 2013-06-09 | 2013-09-04 | 浪潮电子信息产业股份有限公司 | Isomerism mixed calculation multi-platform system using central processing unit (CPU)+graphic processing unit (GPU)+many integrated core (MIC) |
Non-Patent Citations (2)
| Title |
|---|
| SUNPYO HONG等: "An Integrated GPU Power and Performance Model", 《ACM SIGARCH COMPUTER ARCHITECTURE NEWS 》 * |
| 许桢: "关于CPU+GPU异构计算的研究与分析", 《科技信息》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160274636A1 (en) * | 2015-03-16 | 2016-09-22 | Electronics And Telecommunications Research Institute | Gpu power measuring method of heterogeneous multi-core system |
| US10025364B2 (en) * | 2015-03-16 | 2018-07-17 | Electronics And Telecommunications Research Institute | GPU power measuring method of heterogeneous multi-core system |
| CN104778113B (en) * | 2015-04-10 | 2017-11-14 | 四川大学 | A kind of method for correcting power sensor data |
| CN104778113A (en) * | 2015-04-10 | 2015-07-15 | 四川大学 | Method for correcting power sensor data |
| CN105446877A (en) * | 2015-11-04 | 2016-03-30 | 上海聚力传媒技术有限公司 | Method and apparatus for testing power consumption of mobile application |
| CN106686179B (en) * | 2016-12-27 | 2019-06-21 | 东南大学 | A kind of mobile telephone power consumption test macro and method based on test automation |
| CN106686179A (en) * | 2016-12-27 | 2017-05-17 | 东南大学 | Mobile phone power consumption testing system and method based on test automation |
| CN109684144A (en) * | 2018-12-26 | 2019-04-26 | 郑州云海信息技术有限公司 | A kind of method and device of GPU-BOX system testing |
| CN109684144B (en) * | 2018-12-26 | 2021-11-02 | 郑州云海信息技术有限公司 | Method and device for testing GPU-BOX system |
| CN111538636A (en) * | 2020-04-24 | 2020-08-14 | 深圳华锐金融技术股份有限公司 | Computer equipment determination method and device and storage medium |
| CN111538636B (en) * | 2020-04-24 | 2021-11-19 | 深圳华锐金融技术股份有限公司 | Computer equipment determination method and device and storage medium |
| CN111782454A (en) * | 2020-08-05 | 2020-10-16 | 中国人民解放军国防科技大学 | Instruction EPI-based fine-grained GPDSP power consumption testing method, system and medium |
| CN111782454B (en) * | 2020-08-05 | 2023-08-18 | 中国人民解放军国防科技大学 | Fine-grained GPDSP power consumption testing method, system and medium based on instruction EPI |
| CN112150631A (en) * | 2020-09-23 | 2020-12-29 | 浙江大学 | Real-time energy consumption optimization drawing method and device based on neural network |
| CN112150631B (en) * | 2020-09-23 | 2021-09-21 | 浙江大学 | Real-time energy consumption optimization drawing method and device based on neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104461849B (en) | 2017-06-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104461849A (en) | Method for measuring power consumption of CPU (Central Processing Unit) and GPU (Graphics Processing Unit) software on mobile processor | |
| CN104424092B (en) | The page loads duration method of testing and device | |
| Hao et al. | Estimating mobile application energy consumption using program analysis | |
| Pathak et al. | Where is the energy spent inside my app? Fine Grained Energy Accounting on Smartphones with Eprof | |
| Kim et al. | Enhancing online power estimation accuracy for smartphones | |
| Bunse et al. | On the energy consumption of design patterns | |
| Chen et al. | Android app energy efficiency: The impact of language, runtime, compiler, and implementation | |
| Kim et al. | FEPMA: Fine-grained event-driven power meter for android smartphones based on device driver layer event monitoring | |
| Tu et al. | Performance and power profiling for emulated android systems | |
| Yoon et al. | Accurate power modeling of modern mobile application processors | |
| Dietrich et al. | Managing power for closed-source android os games by lightweight graphics instrumentation | |
| Dietrich et al. | Power management using game state detection on android smartphones | |
| Duan et al. | Energy analysis and prediction for applications on smartphones | |
| Dzhagaryan et al. | An environment for automated measurement of energy consumed by mobile and embedded computing devices | |
| WO2017113877A1 (en) | Automatic measurement method for mobile terminal app loading energy consumption based on actual physical measurement | |
| Creus et al. | Optimizing mobile software with built-in power profiling | |
| Zhang | Power, Performance Modeling and Optimization for Mobile System and Applications. | |
| Hung et al. | System-wide profiling and optimization with virtual machines | |
| Dzhagaryan et al. | An Environment for Automated Measuring of Energy Consumed by Android Mobile Devices. | |
| Benmoussa et al. | Open-PEOPLE, a collaborative platform for remote & accurate measurement and evaluation of embedded systems power consumption | |
| Ahmad et al. | Green smartphone GPUs: Optimizing energy consumption using GPUFreq scaling governors | |
| Ahmad et al. | Online cloud-based battery lifetime estimation framework for smartphone devices | |
| Dietich et al. | Estimating the limits of CPU power management for mobile games | |
| Rattagan et al. | Clustering and symbolic regression for power consumption estimation on smartphone hardware subsystems | |
| Almeida et al. | Energy measurement tools for ultrascale computing: A survey |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20190423 Address after: 215123 Linquan Street 399, Dushu Lake Higher Education District, Suzhou Industrial Park, Jiangsu Province Patentee after: Suzhou Institute, Southeast University Address before: No. 2, four archway in Xuanwu District, Nanjing, Jiangsu Patentee before: Southeast University |